CHAPTER7

Dat
aPr
ocessi
ngandAnal
ysi
s

I
ntr
oduct
ion:Over
viewofDat
aPr
ocessi
ngandAnal
ysi
s

Af
terdat
ahavebeencol
l
ect
edf
rom ar
epr
esent
ati
vesampl
eoft
hepopul
ati
onora
census,t
henextst
epi
stoanal
yzet
hem.Dat
aanal
ysi
sisnow r
out
inel
ydonewi
th
sof
twar
epr
ogr
amssuchasSPSS(
stat
ist
icalPackagef
ort
heSoci
alSci
ences)
,Excel
,
andt
hel
i
ke.Thegoalofanyr
esear
chi
stopr
ovi
dei
nfor
mat
ionf
rom r
awdat
a.The
r
awdat
aaf
tercol
l
ect
ionhast
obepr
ocessedandanal
yzedi
nli
newi
tht
hepl
anl
aid
downf
ort
hepur
poseatt
het
imeofdevel
opi
ngt
her
esear
chpl
an.However
,bef
ore
west
artanal
yzi
ngt
hedat
asomepr
eli
minar
yst
epsneedt
obecompl
eted.These
hel
ptoensur
ethatt
hedat
aar
ereasonabl
ygoodandofassur
edqual
i
tyf
orf
urt
her
anal
ysi
s.Thus,t
hecompi
l
eddat
amustbecl
assi
fi
ed,pr
ocessed,anal
yzed,and
i
nter
pret
ed.

7.
1 Dat
aPr
ocessi
ng

Av
erycommonphr
aset
hati
susedbyr
esear
cher
sis“
gar
bagei
n,gar
bageout
.”Thi
s
r
efer
stot
hei
deat
hati
fdat
aiscol
l
ect
edi
mpr
oper
ly,orcodedi
ncor
rect
ly,y
our
r
esul
tsar
e“gar
bage,
”becauset
hati
swhatwasent
eredi
ntot
hedat
asett
obegi
n
wi
th.Ther
efor
e,l
i
keanypar
toft
hebusi
nessr
esear
chpr
ocess,
car
eandat
tent
iont
o
det
ailar
eimpor
tantr
equi
rement
sfordat
aedi
ti
ngandcodi
ng.Because,af
terdat
a
ar
eobt
ainedt
hroughquest
ionnai
res,i
nter
views,obser
vat
ion,ort
hroughsecondar
y
sour
ces,t
heyneedt
obeedi
ted.Thebl
ankr
esponses,i
fany
,hav
etobehandl
edi
n
someway
,thedat
acoded,andacat
egor
izat
ionschemehast
obesetup.Thedat
a
wi
l
lthenhav
etobeent
eredi
n,andsomesof
twar
epr
ogr
am usedt
oanal
yzet
hem.
Aut
hent
ici
tyandr
elev
anceofar
esear
chi
nvest
igat
ioni
sbasedont
heassur
anceof

1
er
ror
-f
reequal
i
tat
iver
eli
abi
l
ityoft
hecol
l
ect
eddat
a.Dat
apr
ocessi
nghast
obe
car
ri
ed outi
n an appr
opr
iat
emanner
.Techni
cal
l
yspeaki
ng,pr
ocessi
ng i
mpl
i
es
edi
ti
ng,codi
ng,cl
assi
fi
cat
ionand t
abul
ati
onofcol
l
ect
ed dat
aso t
hatt
heyar
e
amenabl
etoanal
ysi
s.Dat
apr
ocessi
ngi
mpl
i
esedi
ti
ng,codi
ng,cl
assi
fi
cat
ionand
t
abul
ati
onofcol
l
ect
eddat
asot
hatt
heyar
eamenabl
etoanal
ysi
s.Thef
ir
stst
epi
na
st
ati
sti
cal
inqui
ryi
sthecol
l
ect
ionofdat
arel
ati
ngt
othepr
obl
em underst
udywhena
massofdat
ahasbeengat
her
ed,
iti
snecessar
ytoar
ranget
hemat
eri
ali
nsomesor
t
ofconci
seandl
ogi
calor
der
.Thepr
ocedur
eisr
efer
redt
oascl
assi
fi
cat
ionand
t
abul
ati
onofdat
a.Anot
heri
mpor
tantconv
inci
ng,appeal
i
ngandeasi
l
yunder
stood
met
hodofpr
esent
ingt
hest
ati
sti
caldat
aist
heuseoft
abl
e,di
agr
ams,andgr
aphs.
Gr
aphsr
evealt
het
rends,i
fany
,pr
esenti
nthedat
amor
evi
vi
dlyt
hant
het
abul
ated
numer
icalf
igur
es.Ther
efor
e pr
ocessi
ng compr
ises t
he t
ask ofedi
ti
ng,codi
ng
cl
assi
fi
cat
ionandt
abul
ati
on.Thest
agesar
eher
eunder
;

1.Edi
ti
ng

Edi
ti
ng:i
sapr
ocessofexami
ningt
hecol
l
ect
edr
awdat
a(unedi
tedr
esponsesf
rom
r
espondentexact
lyasi
ndi
cat
edbyt
hatr
espondent
)todet
ecter
ror
sandomi
ssi
on
(
ext
remev
alues)andt
ocor
rectt
hosewhenpossi
ble.Edi
ti
ngi
sthepr
ocessof
checki
ngandadj
ust
ingdat
aforomi
ssi
ons,consi
stency
,andl
egi
bil
i
ty.I
tinv
olv
esa
car
efulscr
uti
nyofcompl
etedquest
ionnai
resori
nter
view orot
hermet
hods.When
t
heedi
tordi
scov
ersapr
obl
em,heorsheadj
ust
sthedat
atomaket
hem mor
e
compl
ete,consi
stent
,orr
eadabl
e.I
nspi
teofacar
efulcol
l
ect
ionbyar
esear
cher
,
t
her
emaybeapossi
bil
i
tyf
orer
ror
sofomi
ssi
onandcommi
ssi
onar
isi
ngandi
tisf
or
t
hispur
poset
hatt
hepr
ocessofedi
ti
ngbecomesnecessar
y.Edi
ti
ngi
sdonet
o
assur
ethatt
hedat
aar
eaccur
ate,consi
stentwi
thot
herf
act
sgat
her
ed,uni
for
mly
ent
ered,ascompl
etedaspossi
bleandhav
ebeenwel
lar
rangedt
ofaci
l
itat
ecodi
ng
andt
abul
ati
on.Edi
ti
ngcanbeei
therf
iel
dedi
ti
ngorI
n-houseedi
tng.Fi
i eldedi
ti
ng:
consi
stofr
evi
ewi
ngoft
her
epor
ti
ngf
ormsbyt
hei
nvest
igat
orf
orcompl
eti
ngwhat
hasbeenwr
it
teni
nabbr
evi
ati
onand/ori
nil
l
egi
blef
orm atat
imeofr
ecor
dingt
he
r
espondent
’sr
esponse.Thi
ssor
tofedi
ti
ngshoul
dbedoneassoonaspossi
ble
af
tert
hei
nter
vieworobser
vat
ion.I
n-houseedi
ti
ng:i
twi
l
ltakepl
aceatt
her
esear
ch

2
of
fi
ce.Thi
sty
peofedi
ti
ngi
mpl
i
est
hatal
lfor
msshoul
dgetat
hor
oughedi
ti
ngbya
si
ngl
eedi
tori
nasmal
lst
udyandbyat
eam ofedi
tor
sincaseofal
argei
nqui
ry.I
ts
obj
ect
ivei
stocor
recter
ror
ssuchasent
ryi
nthewr
ongpl
ace,ent
ryr
ecor
dedi
n
mont
h.Thepost
-f
iel
dedi
ti
ngi
scar
ri
edoutwhenf
iel
dsur
veyi
scompl
etedandal
l
t
hef
ormsofschedul
ehav
ebeencol
l
ect
edt
oget
her
.Thi
sty
peofedi
ti
ngr
equi
res
r
evi
ewofal
lfor
mst
hor
oughl
y.

Theedi
tor
,whi
l
eexami
ningcer
tai
nresponsesoft
her
espondent
s,mayf
indsome
mi
stakesi
nthef
orm ofi
ncompl
ete,v
ague,ori
rr
elev
antanswer
s.Suchi
nconsi
stent
answer
shav
etobeel
i
minat
edorsui
tabl
yandr
easonabl
ymodi
fi
ed.Fur
ther
,ther
e
shoul
dbenor
oom f
orf
ict
it
iousdat
atocr
eepi
n.Hence,t
heedi
torhast
otakecar
e
t
oseet
hatt
heaut
hent
ici
tyoft
hedat
aisi
naper
fectshape.

2.Codi
ng

Codi
ngr
efer
stot
hepr
ocessofassi
gni
ngnumer
icalorot
hersy
mbol
stoanswer
sso
t
hatr
esponsescanbeputi
ntoal
i
mit
ednumberofcat
egor
iesorcl
asses.Such
cl
assesshoul
dbeappr
opr
iat
etot
her
esear
chpr
obl
em underconsi
der
ati
on.They
mustal
sopossesst
hechar
act
eri
sti
cofexhaust
iveness(
i.
e.,t
her
emustbeacl
ass
f
orev
erydat
ait
em)andal
sot
hatofmut
ualexcl
usi
vel
ywhi
chmeanst
hataspeci
fi
c
answercanbepl
acedi
noneandonl
yonecel
linagi
vencat
egor
yset
.Assi
gni
ng
numer
icalsy
mbol
sper
mit
sthet
ransf
erofdat
afr
om quest
ionnai
resf
ormst
oa
comput
er.Ther
emustbeacl
assofev
erydat
ait
ems. Theymustbemut
ual
l
y
excl
usi
ve(
aspeci
fi
canswercanbepl
acedi
noneandonl
yonecel
linagi
ven
cat
egor
yset
).Ther
esponsest
ost
ruct
uredquest
ionssuchas“
yes”or“
no”canbe
st
oredi
nadat
afi
l
ewi
thl
ett
erssuchas“
Y”or“
N.”Al
ter
nat
ivel
y,t
heycanbe
r
epr
esent
edwi
thnumber
s,oneeacht
orepr
esentt
her
espect
ivecat
egor
y.Ther
efor
e,
t
henumber1canbeusedt
orepr
esent“
yes”and2canbeusedt
orepr
esent
“
no.
”I
nst
eadofent
eri
ngt
hewor
dmal
eorf
emal
einr
esponset
oaquest
iont
hatasks
f
ort
hei
dent
if
icat
ionofone’
sgender
,wewoul
dusenumer
iccodes(
e.g.
,0f
ormal
e
and1f
orf
emal
e).

3.Cl
assi
fi
cat
ion

3
Fort
hepur
poseofcl
assi
fi
cat
ionoft
hedat
aint
omeani
ngf
ulandusef
ulcl
asses,
the
pr
ocedur
eofcodi
nghast
obeused.Thi
spr
ocedur
ewoul
dbeadv
ant
ageousi
n
deal
i
ngwi
tht
hedat
ahav
inganumberofchar
act
eri
sti
cs.I
naddi
ti
on,
alar
gev
olume
ofdat
acanbepr
ocessedaccur
atel
y.Mostr
esear
chst
udi
esi
nal
argev
olumeofr
aw
dat
a,whi
ch mustber
educed i
nto homogenousgr
oup.Thi
sfactnecessi
tat
es
cl
assi
fi
cat
ionofdat
awhi
chhappenst
obet
hepr
ocessofar
rangi
ngdat
aingr
oups
orcl
assesont
hebasi
sofcommonchar
act
eri
sti
cs.Dat
acl
assi
fi
cat
ioni
mpl
i
est
he
pr
ocessesofar
rangi
ngdat
aingr
oupsorcl
assesbasedoncommonchar
act
eri
sti
cs.
Dat
ahav
ingcommonchar
act
eri
sti
cspl
acedi
nonecl
assandi
nthi
swayt
heent
ir
e
dat
aar
edi
vi
dedi
ntoanumberofgr
oupsorcl
asses.I
not
herwor
ds,
het
erogeneous
dat
aisdi
vi
dedi
ntosepar
atehomogeneouscl
assesaccor
dingt
ochar
act
eri
sti
cst
hat
exi
stamongstdi
ff
erenti
ndi
vi
dual
s orquant
it
ies const
it
uti
ng t
he dat
a.Thus,
f
undament
all
y cl
assi
fi
cat
ion i
s dependentupon si
mil
ari
ti
es and r
esembl
ances
amongt
hei
temsi
nthedat
a.

Descr
ipt
iveandQuant
it
ati
veCl
assi
fi
cat
ion:Dependi
ngont
hechar
act
eri
sti
csoft
he
dat
a,t
heycan bebr
oadl
ycat
egor
ized i
ntot
wo separ
ateand di
sti
nctgr
oups-
descr
ipt
ive and numer
ical
.Descr
ipt
ive char
act
eri
sti
cs ar
ethose t
hatcan be
descr
ibed i
n wor
ds and ar
e expr
essi
ble i
n qual
i
tat
ive t
erms. Numer
ical
char
act
eri
sti
csar
equant
it
ati
vei
nnat
ure.Fori
nst
ance,l
i
ter
acy
,sex,cast
e,and
r
eli
gionar
edescr
ipt
ivechar
act
eri
sti
cs.Hei
ght
,wei
ght
,age,
income,
andexpendi
tur
e
ar
enumer
ical
l
yexpr
essi
blechar
act
eri
sti
cs.Descr
ipt
iveorqual
i
tat
ivecl
assi
fi
cat
ion
i
s t
ermed cl
assi
fi
cat
ion accor
ding t
o at
tri
but
es. Numer
ical or quant
it
ati
ve
cl
assi
fi
cat
ionofdat
aincer
tai
ncl
assi
nter
val
sist
ermedascl
assi
fi
cat
ioni
nter
msof
cl
asseswi
thcer
tai
nint
erv
als,orcl
assi
fi
cat
ionaccor
dingt
ocl
assi
nter
val
s.For
i
nst
ance,dat
acol
l
ect
edonl
i
ter
acyi
nthecount
rycanbecl
assi
fi
edi
ntot
wodi
sti
nct
cl
asses:l
i
ter
ateandi
l
li
ter
ate.Phenomenal
i
kei
ncome,hei
ght
sandwei
ght
sar
eal
l
quant
it
ati
vel
ymeasur
abl
eanddat
aont
hem canbecl
assi
fi
edi
ntosepar
atecl
ass
i
nter
val
sofuni
for
mlengt
h.Fori
nst
ance,t
hemar
ksobt
ainedbyagr
oupof50
candi
dat
esi
nasubj
ectatanexami
nat
ioncanbecl
assi
fi
edi
ntot
hef
oll
owi
ng
cl
asses:
0-10,
10-
20,
20-
30,
30-
40,
40-
50,
50-
60,
60-
70et
c.

4
4.Tabul
ati
on

Dat
acanbepr
esent
edt
hrought
abul
ati
onorGr
aphi
c/di
agr
ammat
icf
orms

1.Tabul
arPr
esent
ati
onofDat
a

Whenamassofdat
ahasbeenassembl
ed,
itbecomesnecessar
yfort
her
esear
cher
t
oar
ranget
hesamei
nsomeki
ndofconci
seandl
ogi
calor
der
.Thi
spr
ocedur
eis
r
efer
redt
oast
abul
ati
on.Tabul
ati
on:i
sthepr
ocessofar
rangi
nggi
venquant
it
ati
ve
dat
abasedonsi
mil
ari
ti
esandcommonchar
act
eri
sti
csi
ncer
tai
nrowsandcol
umns
soast
opr
esentt
hedat
avi
vi
dlyf
orqui
cki
ntel
l
igi
bil
i
ty,
easycompar
abi
l
ityandv
isual
appeal
.Itr
efer
stot
heor
der
lyar
rangement
sofdat
ainat
abl
eorot
hersummar
y
f
ormat
.Itpr
esent
sresponsesort
heobser
vat
ionsonaquest
ion-
by-
quest
ionori
tem-
by i
tem basi
s and pr
ovi
des t
he mostbasi
cfor
m ofi
nfor
mat
ion.I
ttel
l
sthe
r
esear
cherhowf
requent
lyeachr
esponseoccur
s.Tabul
ati
oni
sessent
ialbecauseof
t
hef
oll
owi
ngr
easons.

 I
tconser
vesspaceandr
educesexpl
anat
oryanddescr
ipt
ivest
atementt
oa
mi
nimum.

 I
tfaci
l
itat
est
hepr
ocessofcompar
ison.

 I
tfaci
l
itat
est
hesummat
ionofi
temsandt
hedet
ect
ionofer
ror
sandomi
ssi
ons.

 I
tpr
ovi
desabasi
sforv
ari
ousst
ati
sti
cal
comput
ati
ons.

Tabul
ati
oncanbedonebyhandorbymechani
calorel
ect
roni
cdev
ices.Thechoi
ce
dependsont
hesi
zeandt
ypeofst
udy
,costconsi
der
ati
ons,t
imepr
essur
esandt
he
av
ail
abi
l
ityoft
abul
ati
ngmachi
nesorcomput
ers.I
nrel
ati
vel
ylar
gei
nqui
ri
es,
wemay
usemechani
calorcomput
ert
abul
ati
oni
fot
herf
act
orsar
efav
orabl
eandnecessar
y
f
aci
l
iti
esar
eav
ail
abl
e.Handt
abul
ati
oni
susual
l
ypr
efer
redi
ncaseofsmal
linqui
ri
es
wher
ethenumberofquest
ionnai
resi
ssmal
landt
heyar
eofr
elat
ivel
yshor
tlengt
h.
Tabul
ati
onmayal
sobecl
assi
fi
edassi
mpl
eandcompl
ext
abul
ati
on.Thef
ormer
t
ypeoft
abul
ati
ongi
vesi
nfor
mat
ionaboutoneormor
egr
oupsofi
ndependent

5
quest
ions,wher
east
hel
att
ert
ypeoft
abul
ati
onshowst
hedi
vi
sionofdat
aint
woor
mor
ecat
egor
iesandassuchi
sdesi
gnedt
ogi
vei
nfor
mat
ionconcer
ningoneor
mor
eset
sofi
nter
-r
elat
edquest
ions.

Component
sofaSt
ati
sti
calTabl
e:Ast
ati
sti
calt
abl
ecompr
isest
abl
enumber
,a
t
it
le,
ast
ubhead,
capt
ionscol
umnsundert
hecapt
ions,
sour
ce.Her
ear
eindet
ail
:

1)Number
:Eacht
abl
eshoul
dbenumber
edsot
hati
tmaybeeasi
l
yident
if
ied.The
numberoft
het
abl
eshoul
dbegi
venatt
het
op,
abov
ethet
it
leoft
het
abl
esot
hat
i
tmayeasi
l
ybenot
iced.I
tcanbegi
veni
noneort
wodi
git
s,separ
atedbyadot
wher
ethef
ir
stdi
gitshowst
hechapt
ernumberandt
heseconddi
gitshowst
he
ser
ial
numberoft
het
abl
eint
hatchapt
er

2)Tabl
eTi
tl
e:i
tindi
cat
est
het
it
leoft
het
abl
eandi
tisr
elat
edt
othespeci
fi
c
quest
ioni
tisgoi
ngt
opr
esenti
nit
.

3)St
ub:Thedesi
gnat
ionsoft
hehor
izont
alr
owsar
ecal
l
edst
uborst
abi
tems.The
st
abi
temsshoul
dbecompl
etedandcl
ear
.Iti
sal
way
sadv
isabl
etocondenset
he
st
abi
temssot
hatt
heymaybewr
it
teni
nonel
i
ne.

4)Capt
ion:Theheadi
ngoft
hecol
umni
scal
l
edcapt
ion.Capt
ionshoul
dbecar
eful
l
y
wor
dedandwr
it
teni
nthecent
eratt
het
opoft
hecol
umn.I
fthedi
ff
erent
col
umnsar
eexpr
essedi
ndi
ff
erentuni
tst
hedef
ini
ti
onoft
heuni
tsshoul
dbe
i
ncl
udedi
nthecapt
ion.

5)Body
:Thebodyoft
het
abl
econt
ainsf
igur
est
hatt
het
abl
eisdesi
gnedt
opr
esent
t
oreader
s.

6)Sour
ce:Thesour
ceoft
hedat
aembodi
edi
nthet
abl
eshoul
dbewr
it
ten.The
sour
cenot
eshoul
dgi
vei
nfor
mat
ionaboutt
hepl
acef
rom whi
chdat
awer
e
obt
ained.I
tiswr
it
tenatt
hebot
tom oft
het
abl
e.

2.Gr
aphi
cordi
agr
ammat
icpr
esent
ati
onofdat
a

Dat
acanbeal
sopr
esent
edt
hroughgr
aphanddi
agr
ammat
icf
orms.Thegr
aph

6
r
efer
stot
he ar
rangementofhor
izont
alas wel
las v
ert
icall
i
nes i
ninch or
cent
imet
er’
sdi
vi
si
ons.I
tisav
erycommonpr
ocedur
etor
epr
esentexper
iment
aldat
a
i
nthef
orm ofagr
aph.Somet
imesagr
aphi
susedt
orepl
aceat
abl
eofdat
ainor
der
t
odr
awat
tent
iont
osi
gni
fi
cantf
eat
uresoft
hedat
athatmaynotber
eadi
l
yappar
ent
.
Forexampl
e,agr
aphof
tenr
eveal
str
endsi
nthedat
a,suchast
hepr
esenceof
maxi
ma,mi
nima,orpoi
ntsofi
nfl
ect
ion.Sev
eralconsi
der
ati
onsar
eimpor
tantf
or
goodgr
aphi
calr
epr
esent
ati
on.Li
ket
abl
e,gr
aphsshoul
dbenumber
edandeach
gr
aphshoul
dhav
eacapt
iont
hatbr
ief
lyandcl
ear
lydescr
ibesi
tscont
ent
.The
dependentv
ari
abl
e,whi
chi
sthemeasur
edv
ari
abl
e,i
snor
mal
l
ytheor
dinat
e(yaxi
s)
andt
hei
ndependentv
ari
abl
e,whi
chi
sthev
ari
abl
eyoucont
rol
,isnor
mal
l
ythe
absci
ssa (
x axi
s).
The mostcommonl
y used gr
aphi
c and di
agr
ammat
ic Dat
a
pr
esent
ati
ont
ool
sar
eli
negr
aphs,hi
stogr
am,andf
requencypol
ygonbarchar
ts,
ogi
ve.
.Et
c.Thehi
stogr
am i
saconv
ent
ionalsol
uti
onf
ort
hedi
spl
ayofi
nter
val
-r
ati
o
dat
a.Hi
stogr
amsar
econst
ruct
edwi
thbar
s(orast
eri
sks)t
hatr
epr
esentdat
aval
ues,
wher
eeachv
alueoccupi
esanequalamountofar
eawi
thi
ntheencl
osedar
ea.A
hi
stogr
am i
sagr
aphi
calwayofshowi
ngaf
requencydi
str
ibut
ioni
nwhi
cht
hehei
ght
ofabarcor
respondst
othef
requencyofacat
egor
y.Li
negr
aphsar
eusef
ult
o
di
spl
aymor
ethanoner
elat
ionshi
pint
hesamepi
ctur
e,f
orexampl
ether
esponset
o
f
ert
il
izeroft
hreedi
ff
erentv
ari
eti
es.Whi
l
ether
eisnogener
alr
ule,
graphswi
thmor
e
t
hanf
ourorf
ivel
i
nest
endt
obecomeconf
usi
ngunl
esst
hel
i
nesar
ewel
lsepar
ated.
Barchar
tsdi
spl
aysi
mpl
eresul
tscl
ear
ly.Theyar
enotgener
all
yusef
ulf
orl
arge
amount
sofst
ruct
uredi
nfor
mat
ion.Si
ncet
hehor
izont
alaxi
srepr
esent
sadi
scr
ete
cat
egor
izat
ion,
ther
eisof
tennoi
nher
entor
dert
othebar
s.

Gener
all
y,whensi
gni
fi
cantamount
sofquant
it
ati
vedat
aar
epr
esent
edi
nar
epor
tor
publ
i
cat
ion,
iti
smostef
fect
ivet
ouset
abl
esand/
orgr
aphs.Tabl
esper
mitt
heact
ual
number
stobeseenmostcl
ear
ly,whi
l
egr
aphsar
esuper
iorf
orshowi
ngt
rendsand
changesi
nthedat
a.

7.
2 Dat
aAnal
ysi
s

The t
erm anal
ysi
sref
erst
othe comput
ati
on ofcer
tai
n measur
es al
ong wi
th

7
sear
chi
ngf
orpat
ter
nsofr
elat
ionshi
pthatexi
stamongdat
a-gr
oups.Dat
aanal
ysi
sis
f
urt
hert
ransf
ormat
ionoft
hepr
ocesseddat
atol
ookf
orpat
ter
nsandr
elat
ions
among dat
a gr
oups.I
nit
s si
mpl
estf
orm,anal
ysi
s may i
nvol
ve det
ermi
ning
consi
stent pat
ter
ns and summar
izi
ng t
he r
elev
ant det
ail
srev
eal
ed i
n t
he
i
nvest
igat
ion.Theanal
ysi
sthatf
ocusononev
ari
abl
edescr
ibedasuni
dimensi
onal
anal
ysi
s,ori
nrespectoft
wov
ari
abl
esdescr
ibedasbi
var
iat
eanal
ysi
sori
nrespect
ofmor
ethant
wov
ari
abl
esdescr
ibedasmul
ti
var
iat
eanal
ysi
s.

Anal
ysi
scanbecl
assi
fi
edasqual
i
tat
iveanal
ysi
sandquant
it
ati
veanal
ysi
s,t
his
cl
assi
fi
cat
ionbasedont
henat
ureoft
hedat
a(numer
ical
/quant
it
ati
veorki
nd/
qual
i
tat
ive)
.Qual
i
tat
iveanal
ysi
sist
heanal
ysi
sofqual
i
tat
ivedat
asuchast
extdat
a
f
rom i
nter
view t
ranscr
ipt
sandopenendedquest
ions.Unl
i
kequant
it
ati
veanal
ysi
s,
whi
chi
sst
ati
sti
csdr
ivenand l
argel
yindependentoft
her
esear
cher
,qual
i
tat
ive
anal
ysi
sisheav
il
ydependentont
her
esear
cher
’sanal
yti
candi
ntegr
ati
veski
l
lsand
per
sonalknowl
edgeoft
hesoci
alcont
extwher
ethedat
aiscol
l
ect
ed.Theemphasi
s
i
nqual
i
tat
iveanal
ysi
sis“
sensemaki
ng”orunder
standi
ngaphenomenon,r
ather
t
hanpr
edi
cti
ngorexpl
aini
ng.A cr
eat
iveandi
nvest
igat
ivemi
ndseti
sneededf
or
qual
i
tat
ive anal
ysi
s,based on a et
hical
l
yenl
i
ght
ened and par
ti
cipant
-i
n-cont
ext
at
ti
tude,and a setofanal
yti
c st
rat
egi
es.Quant
it
ati
ve Anal
ysi
s:numer
ic dat
a
col
l
ect
edi
nar
esear
chpr
ojectcanbeanal
yzedquant
it
ati
vel
yusi
ngst
ati
sti
calt
ool
s
i
ntwodi
ff
erentway
sdescr
ipt
iveanal
ysi
sandi
nfer
ent
ial
anal
ysi
s.St
ati
sti
cal
anal
ysi
s
may r
ange f
rom por
tray
ing a si
mpl
efr
equency di
str
ibut
ion t
o mor
e compl
ex
mul
ti
var
iat
eanal
ysesappr
oaches,
suchasmul
ti
pler
egr
essi
ons.

Onewayofanal
ysi
ngdat
aist
heuseofdescr
ipt
ivest
ati
sti
cssuchasper
cent
age,
measur
esofcent
ralt
endency
,and measur
esofdi
sper
sion.
Descr
ipt
iveanal
ysi
s
r
efer
stot
het
ransf
ormat
ionofr
aw dat
aint
oaf
ormt
hatwi
l
lmaket
hem easyt
o
under
standandi
nter
pret
.Unl
i
kei
nfer
ent
ialst
ati
sti
cs,descr
ipt
ivest
ati
sti
csdonot
gi
ver
esul
tsbey
onddescr
ipt
ion.Descr
ipt
ivest
ati
sti
csar
eusedt
odescr
ibet
hebasi
c
f
eat
uresofdat
a.Summar
ydescr
ipt
ive st
ati
sti
csar
e usual
l
yrepr
esent
ed usi
ng
si
mpl
egr
aphssuchasbargr
aph,
piechar
t,andl
i
negr
aph.Descr
ipt
ivest
ati
sti
cscan
beeasi
l
ycal
cul
atedandgr
aphscanbegener
atedusedMSExcel
,STATA,SPSS,or

8
ot
herst
ati
sti
calpackages.Descr
ipt
iveanal
ysi
sist
heel
ement
aryt
ransf
ormat
ionof
dat
ainawayt
hatdescr
ibest
hebasi
cchar
act
eri
sti
cssuchascent
ralt
endency
,
di
str
ibut
ion,and v
ari
abi
l
ity
.Inf
erent
ialanal
ysi
s:Mostr
esear
cherwi
shes t
o go
bey
ond t
he si
mpl
e cal
cul
ati
on f
requencydi
str
ibut
ion,cal
cul
ati
on av
erages and
di
sper
sion.Theyf
requent
lyconductandseekt
odet
ermi
net
her
elat
ionshi
pbet
ween
v
ari
abl
esandt
estst
ati
sti
calsi
gni
fi
cance.Whent
hepopul
ati
oni
sconsi
sti
ngofmor
e
t
hanonev
ari
abl
e,i
tispossi
blet
omeasur
ether
elat
ionshi
pbet
weent
hem.

7.
3 St
ati
sti
calTechni
quesorTool
sforDat
aAnal
ysi
s

The mostcommon f
orms st
ati
sti
calt
echni
ques ar
efr
equency di
str
ibut
ion or
per
cent
age,measur
ersofcent
ralt
endency
,measur
esofdi
sper
sion,measur
esof
asy
mmet
ry,
cor
rel
ati
on,
andr
egr
essi
on.
1 Fr
equencydi
str
ibut
ion

Whet
hert
he dat
a ar
etabul
ated bycomput
erorbyhand,i
tisusef
ult
o hav
e
per
cent
ages.Tabl
econt
aini
ngper
cent
ageandf
requencydi
str
ibut
ioni
seasi
ert
o
i
nter
pret
.
2 Measur
esofCent
ral
Tendency

Descr
ibi
ngt
hecent
ralt
endencyoft
hedi
str
ibut
ionwi
thmean,medi
an,ormodei
s
anot
herbasi
cfor
m ofdescr
ipt
iveanal
ysi
s.Thesemeasur
esar
emostusef
ulwhen
t
he pur
pose i
stoi
dent
if
yty
picalv
alues ofa v
ari
abl
e ort
he mostcommon
char
act
eri
sti
csofa gr
oup. Measur
esofcent
ralt
endencyar
e al
so known as
st
ati
sti
calav
erage.Themostcommonl
yusedmeasur
eofcent
ralt
endencyi
sthe
mean.Tocomput
ethemean,y
ouaddupal
lthenumber
sanddi
vi
debyhowmany
number
sther
ear
e.I
tisnott
heav
eragenorahal
fwaypoi
nt,
butaki
ndofcent
ert
hat
bal
anceshi
ghnumber
swi
thl
ow number
s.Fort
hisr
eason,i
tsmostof
tenr
epor
ted
al
ong wi
th some si
mpl
e measur
e ofdi
sper
sion,such as t
he r
ange,whi
ch i
s
expr
essedast
hel
owestandhi
ghestnumber
.Themedi
ani
sthenumbert
hatf
all
sin
t
hemi
ddl
eofar
angeofnumber
s.I
tisnott
heav
erage;
iti
sthehal
fwaypoi
nt.Ther
e
ar
eal
way
sjustasmanynumber
sabov
ethemedi
anasbel
ow i
t.I
ncaseswher
e
t
her
eisanev
ensetofnumber
s,y
ouav
eraget
het
womi
ddl
enumber
s.Themedi
an
i
sbestsui
tedf
ordat
athatar
eor
dinal
,orr
anked.I
tisal
sousef
ulwheny
ouhav
e

9
ext
remel
yloworhi
ghscor
es.Themodei
sthemostf
requent
lyoccur
ri
ngnumberi
n
al
i
stofnumber
s.
3 Measur
eofDi
sper
sion

I
tisameasur
ementhowt
hev
alueofani
tem scat
ter
edar
oundt
het
rut
h-v
alueoft
he
av
erage.I
tmeasur
est
hev
ari
ati
onoft
hev
alueofani
tem.Af
teri
dent
if
yingt
he
t
ypi
calv
alueofav
ari
abl
ether
esear
chercanmeasur
ehow t
hev
alueofani
tem i
s
scat
ter
edar
oundt
het
ruev
alueoft
hemean.I
tisameasur
ementofhowf
ari
sthe
v
alueoft
hev
ari
abl
efr
om t
heav
eragev
alue.I
mpor
tantmeasur
esofdi
sper
sionar
e;
r
angemeasur
est
hedi
ff
erencebet
weent
hemax
imum andt
hemi
nimum v
alueoft
he
obser
vedv
ari
abl
e.Meandev
iat
ion:i
tist
heav
eragedi
sper
sionofanobser
vat
ion
ar
oundt
hemeanv
alue(
E(Xi
-X)
/n)
.Var
iancei
smeansquar
edev
iat
ion.I
tmeasur
es
t
hesampl
evar
iabi
l
ity
.St
andar
ddev
iat
ion:t
hesquar
erootofv
ari
anceMeasur
eof
asy
mmet
ry(
skew-
ness)
:whent
hedi
str
ibut
ionofi
temsi
shappent
obeper
fect
ly
sy
mmet
ri
cal
,wet
henhav
eanor
malcur
veandt
her
elat
ingdi
str
ibut
ioni
snor
mal
di
str
ibut
ion.Suchcur
vei
sper
fect
lybel
lshapedcur
vei
nwhi
chcaset
hev
alueof
Mean=Medi
an=ModeSkewnessi
s,t
husameasur
ementofasy
mmet
ryandshows
t
hemanneri
nwhi
cht
hei
temsar
ecl
ust
eredar
oundt
heav
erage.I
nasy
mmet
ri
c
(
nor
maldi
str
ibut
ion)
,thei
temsshowaper
fectbal
anceonei
thersi
deoft
hemode,
buti
naskeweddi
str
ibut
iont
hebal
ancei
sskewedonesi
deordi
stor
ted. The
amountby whi
ch t
he bal
ance exceeds on one-
side measur
es t
he skewness.
Knowl
edgeaboutt
heshapeoft
hedi
str
ibut
ioni
scr
uci
alt
otheuseofst
ati
sti
cal
measur
einr
esear
chanal
ysi
s. Si
ncemostmet
hodsmakespeci
fi
cassumpt
ion
aboutt
henat
ureofdi
str
ibut
ion.Skewnessdescr
ibest
heasy
mmet
ryofadi
str
ibut
ion.
Askeweddi
str
ibut
iont
her
efor
ehasonet
aill
ongert
hant
heot
herdoes.Aposi
ti
vel
y
skeweddi
str
ibut
ionhasal
ongert
ailt
other
ight
.Anegat
ivel
yskeweddi
str
ibut
ion
hasal
ongert
ailt
othel
eft
.Adi
str
ibut
ionwi
thnoskew(
e.g.anor
maldi
str
ibut
ion)i
s
sy
mmet
ri
cal

4 Cor
rel
ati
on

I
sther
eanyassoci
ati
onorcor
rel
ati
onexi
stbet
weent
het
woormor
evar
iabl
e?I
fyes

10
t
henupt
owhatdegr
ee?Thi
swi
l
lbeanswer
edbyt
heuseofcor
rel
ati
ont
echni
que.
Cor
rel
ati
oni
sast
ati
sti
calt
echni
queusedf
oranal
yzi
ngt
hebehav
ioroft
woormor
e
v
ari
abl
es.A cor
rel
ati
oncoef
fi
ciental
way
srangesf
rom negat
iveone(
-1)t
oone
(
1).
Thedi
rect
ionofchangei
sindi
cat
edbypl
usormi
nussi
gns.Thef
ormerr
efer
sto
t
hesy
mpat
het
icmov
ementi
nthesamedi
rect
ionandt
hel
att
eri
nopposi
tedi
rect
ion.
Cor
rel
ati
on:t
hemostcommonl
yusedr
elat
ionalst
ati
sti
ciscor
rel
ati
onandi
tisa
measur
eoft
hest
rengt
hofsomer
elat
ionshi
pbet
weent
wov
ari
abl
es,notcausal
i
ty.
I
nter
pret
ati
onofacor
rel
ati
oncoef
fi
cientdoesnotev
enal
l
ow t
hesl
i
ght
esthi
ntof
causal
i
ty.Themostar
esear
chercansayi
sthatt
hev
ari
abl
esshar
esomet
hingi
n
common;t
hatar
erel
atedi
nsomeway
.Themor
etwot
hingshav
esomet
hingi
n
common,t
hemor
est
rongl
ytheyar
erel
ated.Ther
ecanal
sobenegat
iver
elat
ions,
butt
hei
mpor
tantqual
i
tyofcor
rel
ati
on coef
fi
cient
sisnott
hei
rsi
gn,butt
hei
r
absol
utev
alue.Acor
rel
ati
onof-
0.58i
sst
rongert
hanacor
rel
ati
onof0.
43,ev
en
t
houghwi
tht
hef
ormer
,ther
elat
ionshi
pisnegat
ive.Thef
oll
owi
ngt
abl
eli
stst
he
i
nter
pret
ati
onsf
orv
ari
ouscor
rel
ati
oncoef
fi
cient
s:

0.
8to1.
0 Ver
yst
rong
0.
6to0.
8 St
rong
0.
4to0.
6 Moder
ate
0.
2to0.
4 Weak
.
00t
o0.
2 Ver
yweak
Unl
i
ker
egr
essi
on,
cor
rel
ati
ondoesnotcar
ewhi
chv
ari
abl
eist
hei
ndependentoneor
t
hedependentone,t
her
efor
e,y
oucannoti
nfercausal
i
ty.Resear
cher
sof
tenr
epor
t
t
henamesoft
hev
ari
abl
esi
nsuchsent
ences,r
athert
hanj
ustsay
ing"
onev
ari
abl
e".
Acor
rel
ati
oncoef
fi
ci
entatzer
o,orcl
oset
ozer
o,i
ndi
cat
esnol
i
nearr
elat
ionshi
p.

5. Regr
essi
on

Regr
essi
oni
sther
eanycauseandef
fect(
causal
rel
ati
onshi
p)bet
weent
wov
ari
abl
es
orbet
weenonev
ari
abl
eononesi
deandt
woormor
evar
iabl
esont
heot
hersi
de?
Thi
squest
ioncanbeanswer
edbyt
heuseofr
egr
essi
onanal
ysi
s.I
nregr
essi
on
anal
ysi
sther
ear
etwot
ypesofv
ari
abl
es.Thev
ari
abl
ewhosev
aluei
sinf
luencedori
s

11
t
obepr
edi
ctedi
scal
l
eddependentv
ari
abl
e'andt
hev
ari
abl
e,whi
chi
nfl
uencest
he
v
aluesori
susedf
orpr
edi
cti
oni
scal
l
edi
ndependentv
ari
abl
e.Regr
essi
oni
sthe
det
ermi
nat
ionofast
ati
sti
calr
elat
ionshi
pbet
weent
woormor
evar
iabl
es.I
nsi
mpl
e
r
egr
essi
on,wehav
eonl
ytwov
ari
abl
es,onev
ari
abl
e(def
inedasi
ndependent
)ist
he
causeoft
hebehav
iorofanot
herone(
def
inedasdependentv
ari
abl
e).
Whent
her
ear
e
t
woormor
ethant
woi
ndependentv
ari
abl
es,t
heanal
ysi
sconcer
ningr
elat
ionshi
pis
knownasmul
ti
pler
egr
essi
onsandt
heequat
iondescr
ibi
ngsuchr
elat
ionshi
past
he
mul
ti
pler
egr
essi
onequat
ion.Thi
sanal
ysi
sisadopt
edwhent
her
esear
cherhasone
dependentv
ari
abl
e,whi
chi
spr
esumedt
obeaf
unct
ionoft
woormor
eindependent
v
ari
abl
es.Theobj
ect
iveoft
hisanal
ysi
sist
omakeapr
edi
cti
onaboutt
hedependent
v
ari
abl
ebasedoni
tscov
ari
ancewi
thal
ltheconcer
nedi
ndependentv
ari
abl
es.

Regr
essi
oni
sthecl
osestt
hingt
oest
imat
ingcausal
i
tyi
ndat
aanal
ysi
s,andt
hati
s
becausei
tpr
edi
ctshowmucht
henumber
s"f
it
"apr
oject
edst
rai
ghtl
i
ne.Regr
essi
on
i
soneoft
hev
eryi
mpor
tantst
ati
sti
calt
ool
s,whi
chi
sext
remel
yusedi
nal
mostal
l
sci
ences-
nat
ural
,soci
aland phy
sical
.Iti
s especi
all
y used i
n busi
ness and
Economi
cst
ost
udyt
her
elat
ionshi
pbet
weent
woormor
evar
iabl
est
hatar
erel
ated
causal
l
y,and f
orest
imat
ion ofdemand and suppl
y cur
ves,costf
unct
ions,
pr
oduct
ionandconsumpt
ionf
unct
ions,et
c.Regr
essi
oni
sal
way
sver
yusef
uli
n
model
bui
l
dings.

6.Anal
ysi
soft
imeSer
ies

At
imeser
iesi
sanar
rangementofst
ati
sti
caldat
ainaccor
dancewi
thi
tst
imeof
occur
rence.I
fthev
aluesofaphenomenonar
eobser
vedatdi
ff
erentper
iodsoft
ime,
t
hev
aluessoobt
ainedwi
thshow appr
eci
abl
evar
iat
ions.Exampl
esoft
imeser
ies
ar
ethe ser
ies r
elat
ing t
o pr
ices,pr
oduct
ion and consumpt
ion of v
ari
ous
commodi
ti
esagr
icul
tur
alandi
ndust
ri
alpr
oduct
ion,nat
ionali
ncomeet
c.Anal
ysi
sof
t
imeser
iesi
tusef
uli
nadmi
nist
rat
ion,pl
anni
ngandev
aluat
ionofsoci
o-economi
c
pr
ogr
essaswel
lasf
orr
esear
chi
nvar
iousf
iel
dsofsci
enceandhumani
ti
es.I
tal
so
hel
psi
nanal
ysi
sofaphenomenoni
nter
msoft
heef
fectofv
ari
oust
echnol
ogi
cal
,
economi
candot
herf
act
orsoni
tsbehav
iorov
ert
ime.I
tmakescompar
isont
hati
s
mor
esci
ent
if
icaf
terconsi
der
ingt
hev
ari
ouscomponent
soft
heser
iest
oknowhow

12
t
heyhav
ebehav
edov
eraper
iodoft
ime.

13